A postprocessing method for compensation of scatter and collimator blurring in SPECT: a proof-of-concept study.
A postprocessing method for compensation of scatter and collimator blurring in SPECT: a proof-of-concept study.
复制标题
DOI：
10.2967/jnmt.108.061135
复制
发表时间：
2009-06
影响因子：
1.3
通讯作者：
Zeng GL
中科院分区：
文献类型：
--
作者：
Yan Y;Zeng GL
衰减、散射和模糊是导致SPECT图像退化的三个主要因素。没有对这些退化进行补偿的图像重建会导致对比度降低和定量精度降低。在这个概念验证研究中，我们提出了一种有效的后处理方法来补偿SPECT中的散射和模糊效应。首先仅用衰减校正重建原始图像。然后，在图像域估计一个二维点扩散函数（PSF）来模拟散射和模糊。该空间变化的二维PSF用非对称高斯函数拟合。将估计的二维PSF精度与蒙特卡罗模拟和投影域散射响应函数估计的精度进行了比较。采用进一步模糊和反卷积的方法，利用空间变化的二维PSF恢复图像。通过计算机模拟和模拟实验对该方法进行了验证。初步结果表明，图像质量有所改善，图像对比度和定量精度有所提高，并且该后处理方法是可行的。我们提出了一个概念验证研究的后处理方法，以补偿散射和模糊。我们的研究结果表明，由于该方法易于实现和快速，它是一种有希望的替代最先进的补偿方法。
Attenuation, scatter, and blurring are 3 major contributors to SPECT image degradation. Image reconstruction without compensation for these degradations results in reduced contrast and reduced quantitative accuracy. In this proof-of-concept study, we present an efficient postprocessing method to compensate for the scatter and blurring effect in SPECT. A raw image is first reconstructed with attenuation correction only. Then, a 2-dimensional (2D) point spread function (PSF) in the image domain is estimated to model the scatter and blurring. This spatially variant 2D PSF is fitted with an asymmetric gaussian function. The accuracy of the estimated 2D PSF is compared with that estimated from the Monte Carlo simulations and the scatter response functions in the projection domain. A further-blurring-and-deconvolution method is used to restore images with the spatially variant 2D PSF. The method is tested using computer simulations and a phantom experiment. The preliminary results demonstrate an improvement in image quality, with increased image contrast and quantitative accuracy, and the feasibility of this postprocessing method. We present a proof-of-concept study for a postprocessing method to compensate for scatter and blurring. Our results indicate that the method is a promising alternative to the state-of-the-art compensation methods thanks to its easy and fast implementation.